关键词: 图像识别, 视频监控, 系统架构扩展

Abstract: With the development of image recognition technology, its applications in various fields have been developed rapidly, including but not limited to face recognition applications, prediction and analysis of densely populated areas, perimeter alarm of limited areas, etc. The deep learning model based on image eigenvalue convolution is gradually improved, and the demand for the establishment of the system in the railway industry is increasing day by day. How to make full use of the existing video system to expand the image recognition function so as to optimize the project investment and transformation scope to the greatest extent has become a key and difficult point in the implementation. Through the analysis of the existing video monitoring system and image recognition architecture, this paper puts forward the solutions to expand the image recognition function based on the existing system, compares different schemes, and summarizes the factors that need to be considered in early construction for later expansion, so as to provide reference for the implementation of follow-up projects.

Key words: image recognition, video surveillance, system architecture extension